# Triference-Qwen2 **Repository Path**: haukzero/Triference-Qwen2 ## Basic Information - **Project Name**: Triference-Qwen2 - **Description**: 手写 Triton 融合算子并应用于 qwen2 模型推理 - **Primary Language**: Python - **License**: Not specified - **Default Branch**: master - **Homepage**: None - **GVP Project**: No ## Statistics - **Stars**: 1 - **Forks**: 0 - **Created**: 2025-03-29 - **Last Updated**: 2025-03-29 ## Categories & Tags **Categories**: Uncategorized **Tags**: triton, inference ## README ## Triference Qwen2 使用手写 Triton 融合算子对 PyTorch 函数做修改, 并应用于 Qwen2 大模型推理 融合算子: - RoPE - Attention (forward only) - RMSNorm (forward & backward) 测试模型:Qwen2.5-3B-Instruct